14:59:25 #startmeeting Tor Browser Weekly Meeting 2023-06-20 14:59:25 Meeting started Tue Jun 20 14:59:25 2023 UTC. The chair is richard. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:59:25 Useful Commands: #action #agreed #help #info #idea #link #topic. 14:59:32 hello everyone, the pad per usual: https://pad.riseup.net/p/tor-tbb-keep 14:59:58 and as usual, please tidy up your gitlab boards if they need it 15:02:18 Hi! 15:03:02 ok let's get started 15:03:22 o/ 15:03:40 so I remembered an important bugfix I forgot to submit an MR for last night (tor-bowser#41729) 15:03:48 tor-browser#41729 15:03:53 tor bowser is in another castle 15:04:12 it's just a pref flip an I think we should definitely get it in for 12.5 15:04:37 plus it's mentioned in the release post and as we all know, words can't be changed :< 15:04:56 Can we at least inject it in projects/browser? ^_^; 15:04:58 the release date is tentatively tomorrow, so if we get the MR in (and fix the 12.5.0 -> 12.5 rename we can still be on track if I sign tomorrow morning 15:05:17 Oh, well, that will also need a Firefox rebuild 15:05:34 PieroV: I have a block full of meetings for the next however many hours so I'd be happy to leave it in your capable hands 15:06:01 ack 15:06:22 I'll plan on signing/etc tomorow AM 15:07:23 ok I assigned the issue to you PieroV 15:07:40 Oh, we don't have henry 15:07:46 Matrix bridge problems :( 15:07:51 oh no! 15:07:59 I think they should confirm the thing, though, if they have a Windows VM 15:08:09 Matrix bridge issues? 15:08:17 (seems to be working OK here?) 15:08:27 Unless you can't see my messages either? 15:08:30 Jeremy_Rand_36C3[m]: in the past days 15:08:31 iirc the user and thorin have confirmd ht epref flip 15:08:33 Maybe yesterday 15:08:43 And they haven't re-joined, yet 15:08:46 confirmed the pref flip fixes the issue 15:09:21 and apparently in Win 10 there is a built-in screen reader functionality which demonstrates the issue 15:09:28 ( https://gitlab.torproject.org/tpo/applications/tor-browser/-/issues/41729#note_2910595 ) 15:09:43 thank you zwiebelbot 15:10:06 I'n back home so I have access to win10 again so can prolly test? assuming it really is a built in feature 15:10:15 beyond that, once tor browser 12.5 is out the door I'll do the release prep for mullvad browser 12.5 15:10:24 dan_b: perfect 15:10:43 tho i will be afk after this meeting for ~<2 hours 15:11:00 no worries, it will be a bit for a build to go anyway 15:11:03 richard: so screen reader + tag + change the version to 12.5 + build? 15:11:08 cool 15:11:14 PieroV: exactly 15:11:38 Starting a build already lol 15:11:38 and ping me once it's going so I can kick off my buil as well 15:11:45 hah 15:11:53 ok, donuts 15:11:56 over to you with the weather 15:12:17 are we still on announcements atm? 15:12:26 I think so 15:12:44 cool in that case, please welcome our new product designer jagtalon! 15:12:49 hello, all! 15:12:52 jagtalon: would you like to introduce yourself? 15:13:13 welcome welcome o/ 15:13:32 hello! 15:13:40 Welcome! 15:13:41 o/ welcome weclome :) 15:14:16 yeah! i'm currently living in philadelphia, previously from the philippines. my previous work was at duckduckgo 15:14:31 jagtalon: cool cool, welcome! 15:14:43 thank you for having me! 15:14:54 awesome! welcome 🙂 15:15:11 (it's also my first time doing a text-only meeting i'm kinda digging this) 15:15:11 benvenuto! 15:15:19 the plan is that jag's going to pick up more Tor Browser UX work going forward so I can dedicate more time to the VPN 15:15:48 exciting! 15:15:54 Initially I'll still be doing the UX work for the 13.0 release while jag works on our design systems, which are in dire need of some TLC 15:15:54 yes it's my first day today. doing some onboarding i hope to get to that right after 15:16:14 but I think by the 13.5 release cycle he'll be involved in more sponsor-work/features 15:16:18 Who should we @ on gitlab for UX input? Still donuts for a while, or jagtalon from now on? 15:16:29 henry-x: still me for now please :) 15:16:50 ok! 15:16:54 donuts: I take it the same goes for when I have a UX question on IRC? 15:17:13 Jeremy_Rand_36C3[m]: yep! 15:17:24 cool 15:17:44 I think the design systems work will take a while, but will be a good opportunity for jag to get up to speed on Tor Browser 15:17:49 do y'all have a #tor-ux channel set up we can just pop by and ask and get answers from who evers on? 15:18:11 we do indeed dan_b ^^ 15:18:22 (or maybe an IRC highlight keyword for the UX team?) 15:18:54 Either #tor-ux or #tor-browser-dev are fine :) 15:18:58 although gitlab is preferred 15:19:13 gotcha 15:19:46 okay, over to you pierov? 15:20:02 henry-x: not sure you've read the backlog 15:20:10 But do you still have a Windows VM? 15:20:20 In case could you review a MR with screen reader on Windows? 15:21:38 I'm pretty sure I've used it with a NVDA in the last month on alpha and it was fine, but I can double check 15:22:05 ack. Starting a testbuild soonish on tb-build-05. Thank you! 15:22:22 Apart from that, Tor Browser and Mullvad Browser 115 MRs are ready! 15:22:28 what version do you want testing? 15:22:51 henry-x: we have this issue https://gitlab.torproject.org/tpo/applications/tor-browser/-/issues/41729 15:23:02 And we need to rebuild 12.5 to fix it (and to fix the version number) 15:23:53 ok 15:24:16 Thanks! 15:24:55 ok, donuts: discussion points? 15:25:04 richard: wait 15:25:13 * donuts is waiting 15:25:22 * richard waiting 15:25:25 Again about MRs 15:25:45 https://gitlab.torproject.org/tpo/applications/tor-browser/-/merge_requests/679 15:25:53 https://gitlab.torproject.org/tpo/applications/mullvad-browser/-/merge_requests/62 15:25:58 https://gitlab.torproject.org/tpo/applications/tor-browser-build/-/merge_requests/738 15:26:23 I'd like people to jump on it soonish 15:27:18 "jump" as in...? 15:27:24 Review 15:27:30 I'm planning to review tor-browser-build!738 this week 15:27:37 And switch new developments to 115 15:28:13 boklm: thank you! We'll need the MR for the Stretch update first 15:28:40 (okay, done, I wanted to add the URLs to make sure people found the MRs :)) 15:28:42 pierov: i should be giving trying compiling 115 geckoview a try today 15:29:05 ack, thanks 15:29:17 thank you, then i can test android-components 😄 15:30:33 should I stop waiting now 15:30:39 back 15:30:39 :) 15:30:44 wb richard 15:30:45 sorry 15:30:48 Yes please, go ahead :) 15:30:53 ty pierov ^^ 15:31:25 So there were a few reports on reddit/the forum of downloads failing in 12.0.7 15:31:37 I've created tor-browser#41854 to track the bug 15:31:43 * Jeremy_Rand_36C3[m] ran into that bug as well 15:32:00 however thankfully Alphas new defaults don't seem to be triggering it? So I think we're still good for the release 15:32:23 donuts: there's a pref that you can flip that will fix it 15:32:42 Jeremy_Rand_36C3[m]: interesting, could you add that info to the ticket pls? <3 15:32:54 it's browser.download.enable_spam_prevention 15:32:54 But we flipped it to fix an issue reported in the security review 15:33:23 ma1: oh interesting, and it's also flipped in alpha? 15:33:33 Yeah so I don't fully understand the security issue, but it looks like the "allow" button has no effect in TBB stable 15:33:46 Which might be the "right" thing to fix rather than reverting the pref 15:34:07 donuts, yes, it is a recent "fix" 15:34:12 is there an existing ticket for this pref? 15:34:44 I also don't fully understand how that pref is "supposed" to behave 15:34:52 donuts, looking for it 15:34:58 ty 15:35:12 If the intent is to stop a website from automatically initiating a bunch of downloads without user intervention, that's not what the pref is doing 15:35:13 also lmk where this new "allow" UI exists, because I'm not seeing it in alpha at all 15:35:50 (is it just me or gitlab is not behaving right now?) 15:35:55 Jeremy_Rand_36C3[m]: somehow you appear to be immune to matrix bridge shenanigans 15:36:01 donuts: so, when a download is blocked, there's a button in the Downloads dialog that is supposed to allow the download to continue. IIRC (it's been a week or so since I hit the bug) in TBB stable clicking that Allow button has no effect 15:36:08 ma1: fine for me atm 15:36:20 Maybe it's a problems with strings 15:36:41 If we've changed strings, it could be that an old string has been deleted by mistake, instead of waiting for 12.5 to become stable 15:37:23 donuts: I can try to trigger the bug again but unfortunately it's been a while and I flipped the pref as soon as I realized there was a bug, and it's hard to remember exactly what I did to trigger the bug. I *think* all I did was download a file, and then download another file from the same website? 15:37:28 We turned a bunch of downloads dialogs off by default in Alpha to match Firefox, so that would explain why it's not obvious in alpha 15:37:37 but there aren't any security gains then either 15:38:24 donuts, tor-browser#41764 15:38:30 ma1: ty 15:39:11 idk why this is getting triggered by PDFs then, going by the user reports 15:39:18 something weird is going on here 15:39:31 Ugh, PDFs have their own pref 15:39:33 Let me find it 15:39:35 so does it happen in the current alpha or no? 15:39:38 LOL 15:39:39 this ticket also needed a ~UX tag :3 15:39:47 richard: idk, I don't know how to trigger it 15:39:47 donuts: I vaguely recall that the first download I launched failed (due to a circuit going bad), and then the 2nd download (trying to get the same file) triggered the bug? 15:39:51 browser.download.open_pdf_attachments_inline 15:39:54 But I may be misremembering 15:40:17 This allows PDFs to be opened in the browser, without being downloaded. Maybe it can work as a workaround (and I'd appreciate it :P) 15:40:33 pierov: that may be why the same user reported it was fixed in alpha 15:40:41 but I suspect it's not 15:41:24 Nope, it's something we can flip to prevent PDFs from being downloaded automatically, even though it sounds like a workaround 15:42:08 Oh, hmm 15:42:49 I think the download that triggered it was on a file hosting site (one of MEGA's competitors, can't recall which one). A lot of those sites use HTTP redirects when launching a download, is it possible that's the trigger? 15:43:21 uft the matrix spam is annoying 15:43:54 donuts: am I still immune to whatever Matrix crap is happening? 15:44:02 seem to be! 15:44:14 weird. I don't think my Matrix setup is unusual 15:44:19 anyways 15:44:33 Jeremy_Rand_36C3[m]: it's people joining on IRC, that's why you don't see it 15:44:44 So, major releases invite a lot of attention to Tor Browser and also result in a lot of users upgrading – and I'm not really sure here what the scale of the breakage is here 15:44:46 ^ richard 15:44:57 if you can't find a way to reproduce the bug, ping me on #tor-browser-dev and I'll spend a couple of hours trying to retrace my steps to reproduce it 15:45:11 I've also not tested this at all obvs 15:45:15 sorry, more air-conditioning induced black outs. Are we still talking about the PDF download issue? 15:45:21 ma1: yep :) 15:45:30 I'm hesitant to spend a lot of time on it unless you guys are having trouble reproing it 15:45:48 ma1: I'd like to figure out how to reproduce this in alpha before we can say whether or not it's a problem for the release 15:45:50 (things are hectic this week as I'm mostly occupied shuffling paperwork for a new grant...) 15:46:14 is richard here or has the whole internet imploded? 15:46:17 o/ 15:46:21 lol hello 15:46:22 OK, I'll try to give a shot at it later today. By "alpha" are we still talking of tomorrow's stable? Or 13? 15:46:22 donuts: remember we have candidate binaries 15:46:25 https://tb-build-05.torproject.org/~richard/builds/torbrowser/release/unsigned/12.5.0-build1/ 15:46:32 So, you might want to test here 15:46:34 Instead 15:46:35 pierov: unsigned tho 15:46:46 There's dan_b's script :) 15:47:16 * ma1 guesses we're talking about the 12.5 release candidate(s) 15:47:22 ma1: by alpha I think 12.5 release candidates 15:47:23 I think it would be better if a dev can investigate and post screenshots in the ticket pls 15:47:56 ack 15:48:08 by alpha I meant 12.5a7, because I don't have the RC installed atm 15:48:31 in fact I'm not sure it was available when I replied to the reporter either 15:48:31 well 12.5a7 is p much the same a 12.5 fwiw donuts 15:48:35 yeah :) 15:49:47 ah I need to fix a few "typos" in that script, I'll try that now 15:50:00 So, can someone investigate this today please? To figure out how it can be reproduced, what the bug fix is, and what the new UX is like? 15:50:25 donuts: OK, I'll see if I can repro it later today 15:50:27 May or may not be successful 15:50:28 if the scope of the issue is minor I think the release can continue, but I really don't understand it well enough atm 15:50:42 I mean, it totally broke my workflow until I flipped the pref on stable 15:50:49 Could not download any files 15:51:02 dan_b: can you look into reproing this today? 15:51:07 But I don't know how Alpha may be different 15:51:35 richard: sure I'll give it a try 15:52:17 thx thx please post relevant screenshots for donuts in the ticket 15:52:17 thanks everyone :) 15:52:33 ok, we've 6 minutes before the hour 15:52:37 let's hold off on publishing tomorrow until we know more pls, out of an abundance of caution 15:52:38 is there anything else pressing? 15:53:00 the draft of the release post is here, if folks want to read/review it: https://gitlab.torproject.org/tpo/web/blog/-/merge_requests/199 15:53:13 nothing else on my end, Arthur is still working on the circuit display Namecoin rewrite, and I still need to follow up with the Whonix guys about the circuit display UX for them 15:53:14 please note it's written for a general audience given the reach our releases tend to have 15:53:50 (I would have dealt with the Whonix thing weeks ago but stupid medical issues set me back on everything work-related) 15:54:06 donuts: i'll have a review later today 15:54:25 thanks! 15:54:44 also henry-x, if you could doublecheck the a11y list that would be great thanks <3 15:54:47 I'm not sure how accurate it is 15:54:55 I think reasonably so 15:55:33 ok thanks everyone 15:55:41 #endmeeting